Integration of a feature-based CAD system and an ART1 neural model for GT coding and part family forming

Abstract This paper presents an automated GT (group technology) coding and part family forming system that comprises an ART1 (binary adaptive resonance theory) neural model and a feature-based CAD (computer-aided design) system. The features need to be removed for machining parts are obtained from the feature-based CAD system. After the design process, parts are denoted by binary vectors. The ART1 neural model takes these binary vectors as inputs and forms part families according to the similarities of parts in terms of machining features. Each part and part family are then assigned GT codes according to a customized scheme. The feature-based CAD system and the ART1 neural network system have been implemented in C ++ and C languages, respectively. The working of the whole system is demonstrated with an example.